•Baseline data were acquired from the 8th data release of the Comprehensive Assessment of Neurodegeneration and Dementia (COMPASS-ND)5 alongside 2- to 5- follow-up clinical diagnoses (mean time = 3.01 years, SD = 0.96).

•Olfactory Performance was measured using the Brief Smell Identification Test (BSIT), episodic memory performance was measured using the Brief Visuospatial Learning Test – Revised (BVMT-R) and plasma phosphorylated tau 181 (p-tau 181) concentrations were measured using a single molecule array (Simoa) assay.

•Firth penalized logistic regression models were used to evaluate the independent and additive predictive value of olfactory performance, episodic memory, and plasma p-tau 181. Model discrimination was assessed using receiver operating characteristic (ROC) curve analyses and area under the curve (AUC) metrics.

•No significant within-group differences between converters and non-converters in olfaction, cognition, and plasma p-tau 181, except for within the MCI group.

•CU: None of the predictors were associated with clinically meaningful cognitive decline.

•SCD: Lower baseline BVMT-R scores and higher fluid p-tau 181 levels were associated with a trend towards increased odds of cognitive decline, and both showed moderate discriminative ability for conversion, which improved when combined.

•MCI: Lower baseline BSIT and BVMT-R scores, and higher fluid p-tau 181 were associated with significantly increased odds of cognitive decline. All measures showed moderate to good ability discriminative ability for conversion, which improved when combined.

•Conclusions: These findings suggest that while olfactory identification, alongside cognitive function and biomarkers of AD, predicts disease progression in people with MCI, olfactory changes may not be sufficient in people with SCD to predict future cognitive changes.